Comparison of diameter segment measurement methods For shape functions for single tree volume, biomass and growth estimates Thünen-Institut of Forest Ecosystems Seite Åland, 0 Finland 2.9.-3.9.2015 Source:
Overview Presentation Structure 1. Background 2. Objectives 3. Data & Methods 4. Results 5. Discussion 6. Outlook Future perspectives Comparison of diameter segment measurement methods to obtain stem shape functions for implementation in a flexible self-learning tool for single tree volume, biomass and growth estimates Source: Seite 1
Background Challenges: Slow process for adaptation & renewal of taper equations & thus for the calculation of volume & growth Data originates from historical data of monocultures & even-aged stands Forest management, structure, tree composition & stem form changed Destructive diameter measurement methods for taper equation fitting Benefits & requirements: Own near real time application gives higher flexibility with the opportunity to directly add data and adapt/modify equations and Coefficients IN TIME Data collection of NFI Analysis of data by FVA BW Application of volume functions for NFI & WEHAM Adaption/ modification of equations and coefficients Which leads to time delay Seite 2
Volume calculation in German NFI Volume calculation & assortment structure in the Evaluation of the NFI data is based on the following ideas: The growing stock of sample trees > 7cm BHD is calculated from trees selected by Angle count sampling (counting factor 4) Equations involving the BHD, the upper diameter and tree height, are customized for each single tree measured and integrated to detect the specific stem shape This concept of taper equations (spline functions) is described in KUBLIN et al. (1988) & KUBLIN (2002 Program BDat) Until now, the upper ø had to be measured (in 7m absolute height or 30% relative height of the max height) Basically measurment height of upper diameter could vary but must be far enough from the stem base Seite 3
Basic volume calculation Calculation of volume by splitting tree stem into different segments Conoid ø measurements need no even spacing along stem BUT an appropriate no. of measurements is necessary to represent the real ø changes Calculation of overall volume by adding the volume of tree segments together Improvement of calculations for the individual segments by considering the overall stem shape change of shape/ curvature from base to top Paraboloid/ Cylinder Neiloid Tree Stem Model Seite 4
Objectives Comparison of diameter segment measurement methods to obtain stem shape functions for implementation in a flexible selflearning tool for single tree volume, biomass and growth estimates Overall Objective: Modification of taper equations for volume estimation and assortment structure for the German NFI Methods of first part: Comparison of diameter measurement methods Investigation: Applicability in NFI Accuracy & precision Time efficiency & Costs Source: Seite 5
Diameter Measurement Methods ø measurement is done with dendrometers Study about usability/accuracy from Clark et al. (2000) Dendrometers Contact dendrometers Remote noncontact optical dendrometers Calipers Diameter Tapes Optical calipers Rangefinder Optical fork Pentaprism Bar & Stroud Tele-relaskop Criterion laser Digital Camera BHD upper diameter Seite 6
Diameter Measurement Methods This research could be started in cooperation with a root biomass project conducted at TI Measurements: 1. Optical dendrometers: Trees were marked up to 4 m height Criterion (RD 1000) n=10 Source: Disto (Leica DISTOTM D810 touch) n=8 Digital Camera (Nikon D300) n=16 For all trees 2 images were taken from different angles (also measured: - angle of images & distance to tree, - image specific parameters) Source: Leica DISTO D810 touch, Quick Start, www.disto.com Seite 7
Diameter Measurement Methods 2. Contact dendrometer as reference: Øs of felled tree were measured as reference every meter with Caliper n=16 (3 oaks, 13 birches) BHD was also measured with caliper and additionally girth n=16 Total tree height was measured on felled trees Source: Seite 8
Further Methods 3. Image analysis of terrestrial digital photogrammetric approach based on principles of: image matching/referencing edge or region based segmentation possibly over a seed region growing algorithm polar geometry to transform image informato to a 3D standing tree 4. Least products regression analysis to compare diameter measurement methods data modeling technique to compare measurement methods : observational errors on dependent & independent variables are taken into account Seite 9
Preliminary Results Diameter measurement methods R application to directly compare the different measurements of optical fork dendrometers with the contact dendrometers as referenece (red line) for each tree (with shiny Web application framework for R) Seite 10
Preliminary Results Diameter measurement methods Comparison of the difference of the specific measurements of optical fork dendrometers with contacts dendrometers as reference (red line) further development: comparison of all methods with the Least products regression analysis including the implementation of export functions Seite 11
Discussion Diameter segment measurement methods Criterion Disto Digital Camera Data accquisition quite fast Direct data accquisition? Low cost of equipment Fast Data accquisition With increasing height, marks get not visible No measurement with fog, hard rain, snow Very time consuming measuring every meter Low picture resolution Pixelsize to big for high upper Ø Advanced algorithms needed Light conditions Terrestrial digital photogrammetry seems to be promising BUT need to be tested Advantage to extract additional information from photography later Terrestrial laser scanner data? Seite 12
Outlook TDP Terrestrial digital photogrammetry for diameter measurement? Review of previous studies 0 0 500 1000 1500 2000 500 1000 α Photo 1 Photo 2 1500 2000 2500 3000 Seite 13 Source:
Outlook Possibilities & Methodologically Concept Investigate optical fork method for measurement of whole tree Free open-source R for Statistical Computing offers immense capabilities to investigate, manipulate and analyze data Other opportunities of data acquisition to modify taper functions: Harvester measurements (stanford ) Trunk measurements in saw mills Both not approved measurement methods yet Whatever measurement method is chosen data collection could be continued during control data collection of the Greenhouse Gas Inventory Seite 14
Literature BMELV (2011), Aufnahmeanweisung für die dritte Bundeswaldinventur (BWI 3 ) (2011 2012). Bonn, 2. geänderte Auflage. Clark, Neil; Wynne, Randolph H.; Schmoldt, Daniel L. (2000), A review of past research on dendrometers. Forest Science, 46(4): 570-576 KUBLIN, E. (2002), Verfahrens- und Programmbeschreibung zum erweiterten BWI Unterprogramm BDAT 2.0, Forstliche Versuchs- und Forschungsanstalt Baden-Württemberg, Freiburg Kublin, E. (2003), Einheitliche Beschreibung der Schaftform Methoden und Programme BDATPro. Forstwissenschaftliches Centralblatt, 122: 183 200. doi: 10.1046/j.1439-0337.2003.00183.x KUBLIN, E. (2011), Methodenentwicklung für die dritte Bundeswaldinventur (BWI³). Modul 1: Schaftformmodellierung: BDAT-Aktualisierung. Technischer Bericht, Forstliche Versuchs- und Forschungsanstalt Baden-Württemberg, Freiburg M.W.Taylor (2011), ents The Magazine of the Native Tree Society - Volume 1, Number 12, December 2011 Seite 15
Thank you for your Attention! I am grateful for your constructive criticism and helpful hints. Susann.klatt@ti.bund.de Thünen-Institute of Forest Ecosystems Seite 16 Source: